A data access system has host computers having front-end controllers nFE_SAN connected via a bus or network interconnect to back-end storage controllers nBE_SAN, and physical disk drives connected via network interconnect to the nBE_SANs to provide a distributed, high performance, policy based or dynamically reconfigurable, centrally managed, data storage acceleration system. The hardware and software architectural solutions eliminate BE_SAN controller bottlenecks and improve performance and scalability. In an embodiment, the nBE_SAN (BE_SAN) firmware recognize controller overload conditions, informs Distributed Resource Manager (DRM), and, based on the DRM provided optimal topology information, delegates part of its workload to additional controllers. The nFE_SAN firmware and additional hardware using functionally independent and redundant CPUs and memory that mitigate single points of failure and accelerates write performance. The nFE_SAN and FE_SAN controllers facilitate Converged I/O Interface by simultaneously supporting storage I/O and network traffic.

Techniques described herein are generally related to retrieval of data in computer systems having multi-level caches. The multi-level cache may include at least a first cache and a second cache. The first cache may be configured to receive a request for a cache line. The request may be associated with an instruction executing on a tile of the computer system. A suppression status of the instruction may be determined by a first cache controller to determine whether look-up of the first cache is suppressed based upon the determined suppression status. The request for the cache line may be forwarded to the second cache by the first cache controller after the look-up of the first cache is suppressed.

Systems and methods relate to servicing a demand miss for a cache line in a first cache (e.g., an L1 cache) of a processing system, for example, when none of one or more fill buffers for servicing the demand miss are available. In exemplary aspects, the demand miss is converted to a prefetch operation to prefetch the cache line into a second cache (e.g., an L2 cache), wherein the second cache is a backing storage location for the first cache. Thus, servicing the demand miss is not delayed until a fill buffer becomes available, and once a fill buffer becomes available, the prefetched cache line is returned from the second cache to the available fill buffer.

A method and system for shared virtual memory between a central processing unit (CPU) and a graphics processing unit (GPU) of a computing device are disclosed herein. The method includes allocating a surface within a system memory. A CPU virtual address space may be created, and the surface may be mapped to the CPU virtual address space within a CPU page table. The method also includes creating a GPU virtual address space equivalent to the CPU virtual address space, mapping the surface to the GPU virtual address space within a GPU page table, and pinning the surface.

A printed circuit board assembly (PCBA) for a storage device comprising a non-volatile memory (NVM) and a multi-core processor, wherein a first core of the multi-core processor is devoted to external interface management and a second core of the multi-core processor is devoted to internal data management.

Apparatuses, methods and storage media associated with multiple processor modes execution are described herein. In embodiments, an apparatus may include a processor with a plurality of processor modes, including a first processor mode to address a first address space, and a second processor mode to address a second address space, the second address space including the first address space. The apparatus may further include a signal handler to handle a signal from a kernel, in the first processor mode; and a signal handler wrapper to switch the processor to the second processor mode on delivery of the signal from the kernel, save a current extra context of the second processor mode from the second register file to a user stack, switch the processor back to the first processor mode, then invoke the signal handler to handle the signal. Other embodiments may be described or claimed.

Technologies for managing replica caching in a distributed storage system include a storage manager device. The storage manager device is configured to receive a data write request to store replicas of data. Additionally, the storage manager device is configured to designate one of the replicas as a primary replica, select a first storage node to store the primary replica of the data in a cache storage and at least a second storage node to store a non-primary replica of the data in a non-cache storage. The storage manager device is further configured to include a hint in a first replication request to the first storage node that the data is to be stored in the cache storage of the first storage node as the primary replica. Further, the storage manager device is configured to transmit replication requests to the respective storage nodes. Other embodiments are described and claimed.

A method, software and device for managing a cache service layer of an online solution is described. The online solution includes a database, at least one client, a cache service layer having a plurality of nodes which are interconnected to each other and provide processing and caching power for the cache service layer, and the cache manager. The method comprises reading in a business object from the database; assigning, using a cache manager, the business object to a business object group on a first node of the cache service layer; determining, by the cache manager, the effective probability of cache expiration of the business object group; and setting an expiration time for the business object group based on the determination of the effective probability of cache expiration of the business object group.

Cache lines in a multi-processor computing environment are configurable with a coherency mode. Cache lines in full-line coherency mode are operated or managed with full-line granularity. Cache lines in sub-line coherency mode are operated or managed as sub-cache line portions of a full cache line. Each cache is associated with a directory having a number of directory entries and with a side table having a smaller number of entries. The directory entry for a cache line associates the cache line with a tag and a set of full-line descriptive bits. Creating a side table entry for the cache line places the cache line in sub-line coherency mode. The side table entry associates each of the sub-cache line portions of the cache line with a set of sub-line descriptive bits. Removing the side table entry may return the cache line to full-line coherency mode.

An approach is provided in which a computing system captures content included in a history buffer entry that corresponds to a flush ITAG. The computing system, in turn, uses an execution unit to transmit the content over a results bus to multiple registers and restore at least one of the registers accordingly.